Full article available in PDF version.

The First Silurian carbonate buildup in North America to be correctly identified as a reef was in eastern Wisconsin, in 1862. In the ensuing 125 years, many hundreds of studies were presented on other Silurian reefs in the Michigan Basin and environs. Successive trends in both scientific and economic interests have characterized the quest to learn what riches these reefs may yield. Three periods are recognizable: (1) An early period of discovery, extending until 1926, saw more incorrect ideas to explain the once-mysterious reefs than correct ideas, e.g., ideas of volcanism and upheavals. James Hall, who advanced both incorrect and correct ideas, best typified this period, but T. C. Chamberlin should be credited most for his lasting insight. (2) A middle period of enlightenment, 1927 to 1960, saw a convincing reef proof and a systematic set of biologic reef parameters set forth by (especially) E. R. Cumings, R. R. Shrock, and H. A. Lowenstam. (3) A modern period of integration, 1961 to present, could have been designated as one of proliferation, so numerous were the new reef models and ideas concerning reef geometry, distribution, diagenesis, evaporite relations, deep- versus shallow-water environments, basin-to-shelf differences, cyclicity in deposition, sea-level changes, tectonism, and hydrocarbon accumulation. Many of these ideas conflict; thus, I choose the regionally broad stratigraphic integration that developed as the most significant key to the modern period and the several debates that yet require reckoning against the modern stratigraphic framework. The stratigraphic relations favored in this chapter depart from tradition, but they suggest several kinds of studies that need to be undertaken.

Abstract The Michigan Basin (Fig. 1) is bordered on the west by the Wisconsin Highland and to the north and east of Lake Huron by the Canadian Shield. The Algonquin Arch is a major Precambrian feature trending northeastward in Ontario but becoming almost east-west at the border of Michigan. The northwest Ohio area was a platform during the early part of the Paleozoic, and the Findlay Arch is a post-Silurian structure that developed on this platform. The Findlay Arch plunges northeastward and dies out in the western end of Lake Erie. Contrary to many previously published maps, the Findlay Arch is separated from the Algonquin Arch both in trend and in time of formation. The Kankakee Arch was present by Early Ordovician time as a low-relief feature. Droste and others (1975) have described the detailed geologic history of this area and identified an early to middle Paleozoic feature in the area of the present Kankakee Arch, which they named the Wabash Platform. The Wisconsin Highland is a persistent positive area, which was definitely present by Late Cambrian time. Areas omitted from this study are Manitoulin Island, the Bruce Peninsula, and southwestern Ontario, which are covered in a separate publication by Stott and Aitken (in preparation). Also omitted is the Door Peninsula of Wisconsin, which because of its isolation from the rest of the Michigan Basin is better related to the strata adjacent to the Wisconsin Highland. Northeastern Illinois is included in the chapter on the Illinois Basin. In order to avoid

Abstract In the three-corner area of Indiana, Michigan, and Ohio, much of the 500-ft (150 m) Silurian section younger than the Salamonie Dolomite is a facies of the reef-bearing rocks of the Wabash platform areas of Illinois, Indiana, and Ohio. Rocks of the Salina Group and stratigraphical-ly equivalent rocks of the platform area do not contain salts and anhydrites, but they reflect the Michigan-basin cyclic sedimentation as far south as central Indiana in the form of transgressive-regressive facies within named rock units. No major unconformity is in the section, and the up-dip carbonate rocks probably are lateral equivalents of salts in the basin. Approximate correlation of Wabash-platform rocks in Indiana with units of the Salina in the Michigan basin is: (1) Limberlost Dolomite—lowest part of A unit; (2) Waldron Formation through Louisville Limestone—much of remainder of A unit, especially A-1 carbonate; (3) Wabash Formation—from upper part of A unit (B unit in some areas) through uppermost Salina; (4) Kokomo Limestone Member (Salina Formation)—D unit and possibly younger; and (5) Kenneth Limestone Member (Salina)—probably younger than D. Three reef-start episodes on the platform were coordinated with periods of more normal salinity during late deposition of the Salamonie, late deposition of the Louisville, early in deposition of the Mississinewa, and during deposition of the Kenneth. Some of the earliest reefs aborted during A-unit periods of above-normal salinity, including periods represented by part of the Limberlost and middle Louisville rocks, but many reefs grew during all the time of Salinan cyclic deposition. Even in the northern platform area, where the upper part of Silurian rocks has been eroded, a complete platform buildup that included reefs could have continued to accrete during deposition of the uppermost parts of the Salina Group. These interpretations do not readily favor some current ideas on thick sabkha evaporites, hundreds of feet of drawdown, and near-desiccation in the proto-Michigan basin—nor do they favor regional development of a so-called “Niagaran-Cayugan unconformity.”

Abstract Whether they formed under shallow or deep water, thick geosynclinal sediments were regarded until recently as essential precursors to mountains. Every orogenic belt presumably had evolved stage by stage from geosyncline to mountain, ultimately producing a peripheral accretion to some evergrowing continental craton. Conversely, by implication, thick sediment prisms along any present continental margin inevitably should lead to mountains. These long-standing deterministic generalizations were hardly justified, however, for modern orogenic belts are not consistently located at continental margins, nor do they all contain thick sediments. Moreover, it is impossible to designate any uniquely geosynclinal sediment type. Most geosynclinal sediments are results more than causes of orogenesis; an orthogeosyncline is simply a sediment-filled orogenic belt. The early, strictly uniformitarian sea-floor spreading model for geosynclines was unacceptable because it regarded continental terrace sediment prisms (miogeoclines) formed on passive or nonorogenic continental margins as essential, evolutionary precursors to mountain building—a holdover from the venerable tectonic cycle. But most existing continental terraces are almost 200 million years old and still show practically no tectonic mobility. Genetically, these miogeoclines belong to a different genus than accumulations formed in active orogenic zones. Miogeoclines form on passive trailing edges of diverging continents, whereas orthogeo- synclines form near active leading edges of converging lithosphere plates. Plate tectonics shows how these two genetically distinct sediment prisms may become coincidently crushed together in orogenic belts. Rather than being simple concentric accretions of successive orogenic belts, continents are mosaics of very complexly truncated, overprinted, and even rifted tectonic elements containing haphazard relics of former plate margins. The geosynclinal concept has suffered from an analogue syndrome. Dogmatic generalizations were applied to all cases from an incomplete list of supposed modern analogues, and tectonic environments were confused with sedimentary ones. But intensive marine research over the past two decades has provided many more well- documented possible analogues for the testing of truly actualistic models. Armed with these, as well as with new tectonic insights and new vocabulary, geosynclinal studies can advance from a long descriptive phase to a more genetic one.

Abstract In the 1950's, geosynclinal theory was dominated by the tectogene concept and Marshall Kay's synthesis. These and earlier concepts were derived from field study of tectonized geosynclines on land. In 1959 C. Drake, Maurice Ewing, and G. Sutton, applying the data of marine geophysics, recognized that sedimentary prisms now being laid down along the eastern margin of the United States may represent nascent miogeosynclines and eugeosynclines. They assumed that there is a close parallel with Kay's model and included in their interpretation a shelf-edge basement high that supposedly is equivalent to the tectonic borderland and, also, a toe of sialic crust underlying the continental rise that supposedly makes the rise ensialic. The eugeosyncline then would be elevated eventually to continental level largely by sialization of oceanic crust and without horizontal translation of the prism. Between 1963 and 1967, we have developed what may be called an actualistic concept of geosynclines that is based upon sea-floor spreading and collapsing continental rises. This, too, was based upon Kay's model, except that gross surgery was applied. The seaward half of the miogeosyncline was deleted, as though it never existed and making it a wedge that thickened out, so to speak, like the modern terrace wedge. Also omitted was the tectonic borderland; instead, a continental slope was inserted between the miogeocline and eugeocline. (For simplicity and since none of these sedimentary prisms are really synclinal in form, we prefer the terms miogeocline and eugeocline.) In this model, the miogeoclinal sediments were deposited ensialically on a downflexing continental margin and the eugeoclinal sediments ensimatically on oceanic crust. There seemed to be insufficient reason to equate the shelf-edge basement high with a tectonic borderland or to insert a sialic toe beneath the continental rise. Tectonization was envisioned as the result of underthrusting of the continental margin (sub- duction), which collapsed the continental rise, magmatized it, and inserted allochthonous crust and mantle rock within the eugeocline. Our model is explicitly concerned with the mio-eugeoclinal couplet of the Atlantic type, such as would form marginal to a rift ocean on the trailing edge of a drifting continent, With the rapid development of plate tectonics and especially with the recognition of opening and closing ocean basins, much sophistication has recently been added to geosynclinal theory by J. Dewey, J. Bird, A. Mitchell, H. Reading, W. R. Dickinson, and many others.

Abstract Multiple rapid sea level changes of the past two million years have resulted in major changes in the processes and structures that prevailed for tens of millions of years of pre-Quaternary time during the building of massive midplate continental margins. These changes have hindered interpretations of preQuaternary, or what may be called normal, growth processes and structures to the extent that fallacious concepts of the genetic processes have evolved. We submit that an idealized non-Quaternary midplate continental margin would have a shore zone of prograding beaches but would be devoid of estuaries. Barriers and lagoons would exist primarily in association with delta complexes. Muddy shelf facies would prevail, as no mechanism exists for moving sands out of the shore zone to the shelf in significant quantities. Slopes would be largely smooth and prograding by slow pelagic, hemipelagic, and low-density turbid layer deposition. Erosion by turbidity currents and slumping would be uncommon and largely restricted to the vicinity of a few large long-lived submarine canyons. Axiomatically, the rise would be poorly developed and appear lobate in the form of discrete deep-sea fans off the canyon mouths. Terrigenous turbidites would be mainly confined to these fans.

Abstract Recently, the carbonate belt of the northwestern side of the northern Appalachian Orogen has been interpreted as an early Paleozoic shelf within a continental margin province that evolved synchronously with the This plate tectonic model incorporates results of various studies of the past century, commencing with the original Taconic controversy over the age of Taconic rocks. The last Taconic controversy revolved around the interpretation of Taconic sedimentary rocks that are surrounded by carbonate. Either they are overthrust, as originally suggested by Keith, or they are autochtonous lateral facies within a basin surrounded by the carbonates, as originaly proposed by Dale. The overthrust interpretation was subsequently modified by Ruede- manii to become the Taconic klippen hypothesis that, during the 1960's, has been virtually proven by many workers, but principally by Zen. Kay, as a result of his classic work on the Trenton Group of New York, proposed a miogeosyncline-eugeosyncline (shelf-island arc) couple model for the pre-Middle Ordovician, western Appalachian belt, a concept which was to become the key for future refinement of our understanding of regional stratigraphic-structural relations of the various facies and volcanic assemblages. It provided the basis of the collapsing continental rise model of Dietz and of Dietz and Holden that followed Drake, Ewing, and Sutton's comparison of Kay's model with the present North American Atlantic continental margin. Rodgers then proposed that the southeastward termination of the carbonate shelf was represented by breccia facies that can be observed in the field, that the carbonates and breccias were of a Bahama bank and bank-cdgelike environment, and that this termination was the termination of the Ordovician continental margin. More recently Bird and Dewey incorporated these models into the plate tectonics model of an evolving so-called Appalachian Atlantic Ocean, following Wilson's proposal of a driven-out Proto-Atlantic Ocean and Dewey and Kay's and Commencing in the Ordovician, the Appalachian portion of the enormous sheet of Cambro-Ordovician carbonates of central and eastern North America was involved in the extensive diachronous deformation that formed the Appalachian Orogen. Therefore, reconstruction of the continental margin to its condition before deformation involves the sorting out of bulk stratigraphic-tectonic units, the comparison of their relative chronologies, and the recognition of various sedimentary and structural environments. Using present- day lithosphere plate relations and the analogy of an Atlantic type of continental shelf-rise-abyss facies relationship, a plate tectonics model can be constructed from these various stratigraphic-tectonic elements of the carbonate belt and associated marginal, overlying, and allochthonous rocks that integrates otherwise seemingly diverse and unrelated aspects and provides an actualistic model for the evolution of the North American Ap- The key to reconstructing the pre-Taconian or pre-Middle Ordovician relations is in the stratigraphic- structural assemblages of the Taconian thrust belt. By paleontological considerations of the various structural units or klippen one by one, the chronostratigraphic relations between the carbonate shelf assemblage and the sediments of the thrust sheets can be determined. Essentially, the earliest emplaced thrust sheets contain sediments whose age range matches that of the underlying shelf (autochthon), which was as pointed out a number of years ago by Zen. In addition, the facies of these sediments fit very well the model of shelf- continental rise assemblages that accumulated in a starved environment. The bulk of the initially emplaced klippen, the Giddings Brook Slice, is composed of these offshelf and synchronous sediments. The lowest known fossils in these sediments are in Lower Cambrian orthoquartzite, which matches petrographically The bulk of the overthrust sediments, however, are pre-Lower Cambrian shales and clastics, which are thousands of feet thick and which in the lowest portion are apparently nonmarine graben assemblages containing extrusive basaltic rocks. All these rocks are at least slightly metamorphosed in subgreenschist facies. Additionally, similar facies occur in the autochthon below the Lower Cambrian basal unconformity of the carbonate platform, which also locally contains extrusive basalt and rhyolites. These relationships were Essentially then, with the eastward limit of the carbonate belt being taken as a shelf edge, reconstruction of the overlying structural assemblages indicates a history of late Precambrian continent separation beginning with early horst and graben tectonics and sedimentation, followed by establishment of an Atlantic-type continental margin through to Late Cambrian and Early Ordovician time. Then, ocean closing by subduction along the continental margin converted the margin basement and sediment assemblage to an Andean-like system, diachronously through to Middle to Late Devonian continent-to-continent colli ion. The carbonate belt then lay along the north side of a Himalayan-like mountain system. Its present geographic position is a consequence pf subsequent plate evolution commencing in the Late Triassic as indicated by the Newark Basin assemblage.

Abstract Sediments, processes, and the morphologic profile at intracratonic basin margins commonly arc similar to those of the continental shelf, slope, and rise along open oceanic margins. However, the increased potential for sharp density stratification of intracratonic basin waters and for generation of density currents on surrounding epicontinental shelves can markedly influence depositional and erosional processes on the basin margin or floor and can create distinctive sedimentary features that help to differentiate intracratonic and ocean-margin environments of the geologic record. The mid-Permian outcrops of the Guadalupe Mountains provide excellent examples of both depositional and erosional features of an intracratonic basin margin where sharp density stratification and persistent density currents formed by temperature or salinity differences, rather than bysuspended clay, were important sedimentologic factors. The mid-Permian (Leonardian to early Guadalupian) northwestern margin of the Delaware Basin probably had a normal shelf-slope-rise profile, having several hundred meters of relief, and slopes of a few degrees or less along the basin-margin depositional slope. The exposed 1,000 m of mid-Permian basin-facies strata consist mostly of finely textured dark carbonate rocks, fine-grained sandstones, and siltstones. Carbonate sands and allochthonous carbonate conglomerates and megabreccias derived from the bank or bank margin are locally conspicuous (but minor) interbedded strata. The Leonardian rock units are the contemporaneous bank and basin facies, Victoria Peak dolomites and Bone Springs limestones, respectively. The early Guadalupian rock units are the Cutoff “Shale,” composed mostly of basin-facies limestone, and the overlying Brushy Canyon Formation, composed mostly of detrital sandstone and siltstone. The Cutoff strata lie above and parallel to the basin-sloping unconformity that truncates Leonardian basin-margin deposits. Brushy Canyon strata unconformably onlap both Leonardian and Cutoff strata. The abruptness and position of the Victorio Peak-to-Bone Springs facies change indicate the sharpness and persistence of a euxinic interface along the lower part of the Leonardian basin-margin slope. Currents were generally weak or absent near the interface, but erosion surfaces, some overlain by sheets and channel fills of bank-derived carbonate sands, indicate episodes of higher competence of bottom currents. Intra- and interformational erosion features are more prevalent in Guadalupian strata. Two major erosional phases created unconformities at both the base of the Cutoff and the Brushy Canyon rock units. The unconformities at the basin margin slope 5° to 10° basinward. The lower one truncates about 250 m of Leonardian basin-margin strata, and its carving required appreciable retreat and steepening of the basin-margin depositional slope. The upper unconformity forms the onlap surface for more than 300 m of Brushy Canyon deposits. Several steep-sided, narrow channels as much as 40 m deep incise the sloping unconformity surfaces. Erosion concomitant with sedimentation of basin facies persisted throughout early Guadalupian deposition, and basin- trending channels are especially well displayed in the Brushy Canyon. Brushy Canyon intraformational channel dimensions are substantial, as depths may exceed 25 m, widths 1 km, and lengths many kilometers. Brushy Canyon channels are filled in part by beds of sandstone containing upper flow-regime features that conform to the flatter channel floors and that abut adjacent channel walls. Finely laminated siltstone beds mantle channel floors, walls, and interchannel areas and form the bulk of the Brushy Canyon deposits. The erosive agents that cut both channels and unconformities left clean, smooth contacts but little evidence of their nature. We believe that density currents were the major erosive agent and that all erosion occurred in a relatively deep submarine environment. Evidence for submarine origin includes the basin-facies character of all deposits overlying erosional surfaces, the similarity of small and large scale erosion surfaces, the similarity of the Brushy Canyon erosional features to those of later Guadalupian deposits of established deep-basin origin, and the absence of recognized features of subaerial, vadose, or shallow-marine environments. If sea-level changes were involved, the sea may have been deeper rather than shallower during the carv ing of the major unconformities. The erosional and depositional features of the mid-Permian basin margin are compatible with a basin having sharp density stratification and with frequent spilling of shelf-generated cold or saline density currents down the shelf margin. Denser, bottom-hugging currents carved the channels and probably the unconformities and deposited the coarser grained carbonate and detrital sands. Less dense currents moved partly down the slope and then spread far out over the basin as interflows, creating a rain of finer grained sediment on the deeper basin floor. Density currents may have been frequent, of long duration, and not limited to master channels, thus minimizing proximal-to-distal and fan apex-to-interfan contrasts. Episodic phenomena expectable on any marginal slope, such as debris flows that carried very coarse clasts several kilometers into the basin, or slumps, or perhaps deep wave action, contributed to the sedimentary features. The mid-Permian sedimentary prism of the intracratonic Delaware Basin provides some marked contrasts as well as similarities to sedimentary prisms fronting open ocean basins. In its overall features, it is significant that here is another example from the geologic record for which there appears to he no reasonably close modern analog.

Abstract Aulacogens are long-lived deeply subsiding troughs, at times fault-bounded, that extend at high angles from geosynclines far into adjacent foreland platforms. They are normally located where the geosyncline makes a reentrant angle into the platform. Their fill is contemporaneous with, as thick as, and lithologically similar to the foreland sedimentary wedge of the geosyncline but in addition has periodically erupted alkalic basalt and fanglomerate. Although many aulacogens have suffered mild compressional deformation, tectonic movement within them is mainly vertical; large-scale horizontal translations are rare. Aulacogens are known throughout the Proterozoic and Phanerozoic, and incipient aulacogens occur at reentrants on modern continental margins. The 1700-to-2200-million-year-old Athapuscow Aulacogen of Great Slave Lake began as a deeply subsiding transverse graben during the early miogeoclinal stage of the Coronation Geosyncline. During the orogenic stage of the geosyncline, the aulacogen became a broader downwarp that received abnormally thick exogeosynclinal sediments from the orogenic belt. The aulacogen was compressed mildly, prior to a final stage involving transcurrent faulting, one-sided uplift, and continental fanglomerate sedimentation. The aulacogen is distinguished from the foreland sedimentary wedge of the geosyncline by having paleocurrents parallel rather than transverse to its structural trend, by having high-angle faults rather than low-angle thrusts, by its alkalic basalt volcanism, and by the lack of metamorphism. It is hypothesized that deep-mantle convective plumes produce three-armed radial rift systems (rrr triple junctions) in continents stationary with respect to the plumes. If only two of the arms spread to produce an ocean basin, the third remains as an abandoned rift extending into the continental interior from a reentrant on the new continental margin. For example, the Benue Trough, located in the Gulf of Guinea reentrant on the west coast of Africa, may be such an abandoned rift arm formed during the Cretaceous period at the time of initial rifting of Africa and South America. Inasmuch as new continental margins are predestined to become geosynclines, such abandoned rift arms are juvenile aulacogens. In this model, aulacogens and geosynclines have a common origin but differ in the extent of rifting.

Abstract Dispersal of sediment across a submarine fan is controlled by a distributary system of migrating fan valleys, which are generally contiguous with a feeding canyon that provides a point source for the sediments moving onto the sea floor. Under conditions of fan growth, one active, leveed fan valley on the upper fan leads to a suprafan, a depositional bulge with an irregular surface that is probably the site of most rapid aggradation on the fan. Below the suprafan, the fan surface is relatively smooth and apparently free of distributary channels. Fan-building processes are sensitive to changes in the rate of sediment supply, grain-size distribution within the sediments, and tectonic disturbances within fan and source areas, but they are relatively insensitive to the shape of the depositional basin or to scale factors (ultimate size of the deposit). Specific deep-sea fans included in this discussion are listed in order of decreasing radial dimensions: Bengal Fan (3000 km), Monterey Fan (300 km), La Jolla Fan (80 km), San Lucas Fan (60 km), Navy Fan (60 km), Coronado Fan (50 km), and a small fan in western Lake Superior (5 km). Submarine canyons allow coarse sediment to bypass shelf environments. To remain active, a canyon must maintain its head in or near the surf zone to intercept the littoral drift; a rapid rise in sea level can cut off its major source of sediment. Ascension Canyon, one of several canyons leading to the Monterey Fan, is now relatively inactive as its head lies near the outer shelf. In the California Continental Borderland, Coronado Canyon likewise has been inactive during the present interglacial period, and reflection-profiling data suggest that it was probably inactive during earlier interglacial periods. The rapid sea-level fluctuations of greater than 100 m during the Pleistocene Epoch may be considered extreme, but similar effects may occur when submarine canyons experience large, apparent sea-level changes due to tectonic tilting. In either event, sudden changes affecting sediment supply to canyon heads cause a marked shift in the locus of turbidite deposition. Depositional sites on a deep-sea fan also may change in response to fluctuating sea levels even when the canyon continues to receive sediment. If headward erosion within a canyon head keeps pace with transgression, incision of the fan valley on the upper and middle fan results; the La Jolla fan valley has cut across the entire fan area, and deposition is now confined to a suprafan farther down the San Diego Trough.

Abstract Many flysch, turbidite, fluxoturbidite, and grain-flow sequences from ancient geosynclines probably have been deposited in deep-sea fans adjacent to continental margins. We have obtained stratigraphic and sedi- mentologic criteria for recognizing ancient fan deposits by comparing the Astoria Fan, a large open-ocean fan off the coast of northern Oregon, with the Eocene Butano Sandstone, an ancient continental borderland fan deposit of similar size in the Santa Cruz Mountains, California. Deep-sea fan deposits consist of channel and interchannel facies. Both facies change significantly downfan and laterally across the fan as a result of decreasing current velocities during each turbidity current and as a result of the lateral migrations of channels through time. Geologic mapping reveals thick-bedded, coarser grained, and lens-shaped channel deposits intermixed with thin-bedded and finer grained interchannel deposits. Sand-shale ratios are high within channels and low within upper fan interchannel and distal fan areas. The coarsest grained and thickest bedded gravels and sands are deposited by channelized sediment gravity flows in the submarine canyons and upper fan valleys. These ungraded, poorly sorted, and massive channel sediments change by midfan to thinner bedded, finer grained, vertically graded, and better sorted turbidite sands that contain sedimentary structures in Bouma sequences. Turbidites in interchannel areas are formed by overbank spilling and consist of thin-bedded fine-grained sands and silts characterized by Bouma cde and de sequences. The delineation of fan margins and paleogeography is aided by the lateral and downfan changes in thickness, texture, composition, and paleocurrent directions of fan sediments. High contents of terrigenous debris are present in the sand fractions of hemipelagic muds deposited near the continental margin, and this may help delimit the shoreward boundary of the fan; in contrast, gradation to high contents of pelagic material indicates the direction of the seaward edge of the fan. Radially oriented paleocurrent patterns define the fan apex but are typically complex because of lateral overflow out of and away from channels and because of meandering and lateral shifting of fan channels. The outlining of fan geometry and major channels also is assisted by the decrease of the maximum clast size and of thickness of turbidite beds both downfan and laterally from channels. Variations in morphology, stratigraphy, sedimentary facies patterns, grain-size distribution, sediment composition, and sediment dispersal patterns help identify fans from different geosynclinal settings such as restricted borderland or marginal sea basins, open ocean continental rises, and deep-sea trenches.

Abstract Examination of a number of ancient turbidite basins in the Alpine geosynclinal chains of the circum- Mediterranean region supports the assumption that many of the sediments therein were deposited in deep-sea fan environments. Sand—body geometry and vertical sequence analysis provide criteria for detecting associations of inner, middle, and outer fan facies in these turbidite sequences. Examples of these three main facies associations are reported from selected Tertiary geosynclinal tur- bidites occurring in the northern Apennines (Ranzano Sandstone and Bobbio Formation) and on the island of Rhodes, Aegean Sea (Messanagros Sandstone). The proposed depositional model based on these studies is not unlike models of certain deltaic systems and emphasizes progradational, aggradational, and recessional events of turbidite sedimentation in complexes of ancient deep-sea fans. Middle fan deposits commonly show thinning and (or) fining upward cycles developed within channel-fill sequences. Such cycles are readily comparable to those of abandoned fluvial channels that are also filled with similar sequences in delta-plain environments. In both cases, channel and interchannel areas indicate prevalent vertical accretion. Active deltaic outbuilding is expressed typically by stream-mouth bars whose progradational character is shown by the occurrence of thickening and (or) coarsening upward cycles. Detailed inspection of several northern Apennine turbidite formations has shown that sandstone bodies, closely exhibiting such a progradational pattern, are extremely widespread. These turbidite sediments are here interpreted as outer fan deposits, and are thought to be responsible for deep-sea fan growth in most ancient basins.

Abstract Many modern submarine canyons and deep-sea fans originated in pre-Pleistocene time. Similar submarine canyons, fans, and fan valleys are found in the geological record certainly as far back as the Precatnbrian. Criteria for recognizing ancient submarine channels include: (1) proved or inferred size comparable to modern canyons and fan valleys; (2) comparable geometry (e.g., high axial gradients diminishing seaward and steep wall slopes, some of which become vertical or overhanging) ; (3) similar locations ( or submarine canyons) between shallow-marine (shelf) and deep-marine (basin) environments and (for fan valleys) incision into inferred deep-sea fans at the lower ends of canyons; (4) similarities in lithologies, grain sizes, and primary structures of the fills and their variations along the length and width of the canyons or valleys; (5) similarities in the observed or deduced processes of fast cutting and filling, together with clean-cut channel contacts, concave upward form, minor channels at the base of or within channel fill, comoaction effects, and partial flushing out; and (6) similarities of multiple origin of faunas within the fill (indigenous, swept in from surrounding shelf areas, or derived from the canyon walls). These criteria emerge from a study of all available data on ancient submarine canyons and fan valleys from 32 areas. The information is grouped, tabulated, and discussed under eight stratigraphic and geographic headings: (1) Lower Paleozoic of the Caledonian-Appalachian Geosyncline, (2) Carboniferous of the Pennine Basin, (3) Upper Paleozoic of the Variscan Geosyncline, (4) Permian of the Delaware Basin, (5) Mesozoi and Tertiary of the Tethyan Geosyncline, (6) Mesozoic and Tertiary of California, (7) Tertiary of the Gulf Coast Basin, and (8) other areas. A study of these and future examples of ancient submarine canyons and fan valleys is important as an aid both in reconstructing the continental slopes and rises of geosynclinal and other basins and in the search for possible oil and gas traps. Further intensive studies of transitional areas between shelf and basin facies should reveal many more examples of ancient canyons and fan valleys.

Abstract Many submarine canyons and channels have been described in modern seas, but little conclusive evidence about their existence has been found in ancient rock series. Like the modern submarine canyons, the ancient counterparts were situated on unstable edges of continents. During their further geologic history, they usually underwent extensive tectonic and erosional destruction or were buried below younger sediments. The western part of the Carpathian Flysch Belt in the territory of Czechoslovakia in central Europe is one of the convenient places where the critical zone between the platform and the former geosynclinal trough can be studied. Among the most interesting contributions of this investigation is the discovery of two large buried depressions described as Nesvacilka (N) and Vranovice (V) grabens (fig. 1). Their existence was proved both by geophysical measurements and drilling operations. The depressions, traditionally regarded as tectonic structures, however, show many similarities with modern submarine canyons. The depressions are cut in Paleozoic and Mesozoic carbonate rocks covering the crystalline complexes of the Bohemian Massif. They are filled with Eocene and Oligocene deposits and overlain by Neogene sequences of the Carpathian foredeep. These autochthonous formations deposited on marginal sectors of the platform dip below the Carpathian flysch nappes that comprise Cretaceous and Paleogene miogeosynclinal scries (fig. 1). The longitudinal axes of the depressions are oriented in a northwest-southeast direction perpendicular to the margin of the platform. The canyons probably join each other farther dovvndip to form one channel system not unlike the Scripps and La Jolla Canyons along the coast of California. At their upper ends, the depressions are surrounded by steep, high walls resembling the heads of modern canyons. The thickness of the sedimentary fill near the head of the Nesvacilka Depression exceeds one thousand meters. Even though this thickness does not wholly correspond to the original depth, it indicates the rugged relief of the structure. Downward the canyons become shallower and their walls less steep. The measured width of the depressions varies from about 2 km at their heads to as much as 7 km in their distal parts. Both structures have been followed for a distance of about 25 km, their further courses being hidden beneath 4- to 7-km-thick flvsch and molasse sequences. The Eocene and Oligocene sediments filling the depressions are composed predominantly of dark-brown calcareous silty shales rich in organic matter. They contain abundant planktonic microfauna, proving the marine origin of these sediments. The shales are intercalated with laminae and thin beds of siltstones and fine sandstones. The lamination is a common structure present elsewhere. Scour-and-fill structures of small size found in many sandstone beds indicate the activity of strong, erosional, bottom-seeking currents sweeping through the canyons. No turbidites showing the characteristic succession of internal structures have been identified, but, because of the small number of drill cores available, their presence cannot be excluded. Thick beds of massive sandstones and boulder conglomerates were found at the bottom of the Nesvacilka Depression. Though the oldest deposits found inside the canyons are of late Eocene age, such huge structures should have originated much earlier. Redeposition of Cretaceous and Paleocene micro fauna indicates that the canyons could have been active during the entire time of existence of the flysch miogeosyncline from Cretaceous to Oligocene. The early stage of canyon development was characterized by erosion and transportation of material but was followed during Oligocene time by sedimentation inside the canyons, which brought about the end of their development. The origin of submarine canyons has not been explained satisfactorily, though several hypotheses have been submitted. The Carpathian canyons, oriented parallel to the fault system of the platform, are believed to be of combined tectonic and erosional origin. They intersected the shelf and entered the former geosynclinal flysch trough, which, considering the morphology of the canyons, must have been at least one thousand meters deep. There is an open question what role the canyons played in geosynclinal sedimentation. According to the commonly accepted paleogeographical concept, the flysch trough of the Western Carpathians was supplied predominantly from internal sources (cordilleras), while the platform yielded mostly only fine pelitic material. The distribution of sandy and shaly facies and the composition of clastics apparently support this hypothesis. Modern oceanographic explorations show, however, that canyon-derived sediments can be transported long- distances before finally being deposited. In the San Diego Trough, for example, the coarser sediments accumulate at the distant oceanic side of the basin far from the mouth of supplying canyons along at the coast of California (Shepard, Dill, and Rad, 1969). Also, in the Carpathian Trough the sandy facies were not necessarily related to some nearby sources such as cordilleras. The potential existence of large submarine canyons at the continental side of the Carpathian flysch trough, serving as conduits for flysch sediments, therefore must be taken into consideration in any paleogeographical reconstruction.

Abstract The earth's great mountain belts contain sediments and volcanic complexes of kinds that are well known in the stable cratonic regions and of kinds that are exposed only (or almost only) in the mountain belts. The latter kinds include ophiolites, great sequences of pillow basalts and (or) andesites, bedded, generally radio- larian cherts, uniform shale sequences containing pelagic biotas and representing long time spans, pelagic carbonate sequences, great bodies of turbidite flysch, olistostromes, and metamorphic rocks that are not dealt with here. The sediments are commonly red colored. For a century, explanations for these rocks have been sought along two different lines (paradigms) : In one they are viewed as the products of mobile belts and orogeny--specifically, as the contents of a special generative trough or complex of troughs, the eugeosyncline of Hans Stille and of Marshall Kay. In the other, which we may term the oceanic model and which was stated most clearly by R. S. Dietz and J. C Holden, they are viewed as the normal rocks of the oceans and oceanic margins brought to the mobile belts by plate subduction and incorporated into the continental margins by orogeny. Progressive exploration of the oceans, first by dredging and coring and more lately by drilling, has shown that most of these rocks indeed correspond closely to the rocks of the great ocean floors. This seems to be particularly true when sediments of the same ages are compared. Only the great andesite sequences and olistostromes remain as rock types linked to mobile belts or convergent plate margins. A good case can be made for a distant oceanic derivation of some ophiolite-sediment sequences in mountains (e.g., Elba), and therewith the eugeosynclinal model seems to have outlived its usefulness. At the same time, a model substituting the main ocean basins for the eugeosyncline is even less realistic. Through the work of D. E. Karig and others, confirmed by legs 6 and 7 of the Deep Sea Drilling Project, it has become essentially certain that juvenile oceanic or semioceanic crust overlain by oceanic type sediment can grow in mobile belts to produce the interarc and marginal or small-ocean basins. Yet other sites of crustal growth are likely (e.g., Canary Islands). The ophiolites and sediments of such sites have a better chance of unmetamorphosed incorporation into mountains than does the main sea floor. The ophiolite complexes of Cyprus (Mamonia, Troodos) and of Kandahar, Afghanistan, should be viewed in this light. Furthermore, the odd sediments of the mountain belts commonly form parts of otherwise normal or mio- geosynclinal successions (e.g., Martinsburg flysch, Jurassic sediments of Northern Limestone Alps), and the great andesite sequences occur either on oceanic crust (Antilles) or on sial (Andes). Olistostromes are most likely trench generated. Thus, in principle, the odd rocks of the mountain belts represent a wide range of generative settings, including crust-generating ocean ridges and their flanks, abyssal plains, trenches, volcanic arcs, tectonic welts, interarc basins, and the downwarped edges of continents. Tectonic and paleogeographic reconstructions will remain extremely uncertain until we have learned to recognize these constituents more clearly.

Abstract The main Alpine-Mediterranean Mesozoic lithofacies, excluding flysch, are outlined and placed in their paleogeographic setting within the broader context of an evolving ocean basin. In the external (ophiolite-free) zones of the Alpine-Mediterranean orogen, Mesozoic pelagic facies almost invariably overlie kilometers-thick successions of Bahamian-type platform carbonates. Wherever the basement of these platform carbonates is exposed, it is continental, comprising low- to high-grade metamorphic rocks and granites. We suggest that the pelagic sediments of these zones were deposited on a deeply submerged continental margin of the Atlantic type. Palinspastic reconstructions of the central Alpine-Mediterranean area place the depositional setting of most of these pelagic facies on the southern continental margin of the Tethys. In this area supply of clastics and organic matter was minimal, thus encouraging pelagic conditions. The earliest pelagic sediments of the Alpine-Mediterranean region are of Triassic age and comprise gray and red limestones or cherts that commonly are associated with volcanics. These sediments were deposited in embayments and basins between extensive carbonate platforms and reefs. During the Liassic Epoch, a phase of block faulting, probably related to rifting in the oceanic Tethys, destroyed many of these shallow-water sites, and pelagic conditions became more widespread. During the Jurassic Period a basin-swell morphology was produced by irregular subsidence of the different blocks. On submarine highs, or seamounts, the following stratigraphically condensed facies were developed: pisolitic ironstones, red biomicrites containing ferro- maganese nodules and crusts, crinoid-pelagic bivalve-gastropod-ammonite biosparites, pelagic pelmicrites and micro-oncolitic sparites, and certain red, fine-grained, nodular limestones. In the neighboring basins, more expanded successions containing slumped blocks and turbidites accumulated; the basinal facies were developed as red, more clay-rich, nodular limestones, gray limestone-marl interbeds, radiolarian cherts, and white nannofossil limestones. The Cretaceous Period saw a smoothing of submarine topography and a general deepening of the water as the continental margin continued to founder. Deposition of varicolored marls and red and white coccolith limestones was widespread. True ocean-floor lithofacies are represented by those rocks associated with, or lying upon, ophiolites. In the central Mediterranean area they comprise ophicalcites, umbers, radiolarites, white nannofossil limestones, and black shales. Their age is Jurassic and Cretaceous. In the western central Atlantic pelagic facies of Late Jurassic and Cretaceous ages occur. These facies resemble both the continental margin and ocean-crust lithologies of the Alpine-Mediterranean Tethys. A section through the Mesozoic portion of this undeformed continental margin and ocean-basin complex comprising the Bahamas, the inter-platform straits, and oceanic realm illustrates a paleogeographic arrangement that strongly resembles the reconstructed section for the Alpine-Mediterranean Tethys. This resemblance illustrates the parallel evolution of these two now widely separated areas, so that they both can be considered as representatives of an east-west Mesozoic seaway, or Tethyan realm, that stretched from the Caribbean to Indonesia.

Abstract The main part of the Japanese Islands consists of the middle and late Paleozoic and the early Mesozoic eugeosynclinal sequences containing voluminous volcanic beds. The change through time and space of the nature of the volcanic rocks and the lithologic assemblages of the Paleozoic sequences suggest that they were formed in an ancient marginal or interarc basin and on the outer flank of a migrating ridge. The basin is presumed to have been initiated by a volcanotectonic rift zone that was associated with silicic tuffs and flows and coralline limestones during Middle Silurian to Middle Devonian time. This was followed by the crustal extension that resulted in the separation of the Kurosegawa-Abukuma-South Kitakami insular ridge from the Hida belt to form a marginal basin during Late Devonian to Middle Permian time. Thick eugeosynclinal deposits, including mafic volcancics and chert, accumulated concomitantly in the spreading basins. In the early stage (Late Devonian to Early Carboniferous) of basin spreading, alkaline basalt was produced in association with a large amount of gabbroic rocks and serpentinite, which are assumed to be an oceanic basin crust created along the basin axes. In the later stage (Late Carboniferous to middle Permian) flow rocks and pyroclastics of mainly alkaline, partly tholeütic basalt dominated. They are conformably intercalated in clastic sediments, and they are arranged in an alternating en echelon pattern of a few to several linear volcanic zones. Two zones of ophiolitic assemblages, the Mikabu Green Rocks (Early Permian) and the Yakuno Basic Rocks (Middle Permian), are interpreted to have occurred as short-lived volcanic rises within the marginal basin. The Triassic and Cretaceous mafic volcanic zones respectively were remarkably stepped southward along with the shifting of the main sedimentary basin to the outer side. This shifting was almost coeval with the orogenic events in the main sedimentary basin during the preceding Paleozoic ages. The mafic volcanic rocks are commonly accompanied by bedded chert, and a rhythmic interlayering of tuff and chert is most characteristic. The cherts consist essentially of Radiolaria, some sponge spicules, their fragments, and a small amount of aphanitic material, probably volcanic clay. Their sedimentary features and intimate association with tuff suggest that chert beds accumulated rapidly because of an enormous supply of siliceous organisms. These organisms perished episodically in great masses when explosive submarine eruptions at depths not greater than 500 m resulted in heated water columns rising to the surface and spreading detrimentally among the plankton populations.

Abstract The Paleozoic stratigraphic-structural belts of the Cordilleran foldbelt from Alaska to Nevada and California can be explained by a succession of marginal ocean basins opening and closing behind volcanic arcs. Cambrian through Devonian rocks of the foldbelt are divided into (1) a carbonate rock and quartzite belt (miogeosyncline), a continental shelf and upper continental slope deposit, (2) a graptolitic shale and chert belt (inner part of eugeosyncline), a marginal ocean-basin deposit developed behind a volcanic arc, and (3) a vulcanic rock and graywacke belt (outer part of eugeosyncline), a volcanic arc deposit. The volcanic rock and graywacke belt had a long history of volcanic island development, which began in northern California no later than the Late Ordovician Epoch and in southeastern Alaska and western Washington during the Precambrian or Cambrian Period. Initially, rocks of this belt formed a primitive crust on oceanic basement and through a long process of volcanism, erosion and sedimentation, metamorphism, and plutonism developed a crust of more continental character. In this interpretation, the outer eugeosynclinal rocks are not continental fragments of Asia or parts of oceanic features of unknown provenance that coll'ded with North America. Instead, they are parautochthonous deposits of volcanic island chains and interarc basins developed outward from the North American continental margin as it faced a proto-Pacific basin. Deep-sea pelagic deposits, condensed sections of graptolitic shale and chert, apparently accumulated on basaltic basement in newly formed marginal ocean basins behind the volcanic chains as they migrated from the continent. In the Late Devonian and Mississippian, these marginal ocean basins closed, perhaps by collision with the frontal volcanic arcs, and the deep-sea pelagic sediments were thrust eastward onto the continental shelf, uplifted, and eroded to provide thick wedges of coarse clastic sediments. In the Pennsylvanian and Permian Periods, the collapsed earlier Paleozoic marginal ocean basins were rifted, and a new system of basins, floored apparently by new crust, developed behind segments of the older volcanic arc complexes. This late Paleozoic basin in the latitude of Nevada was closing during the Late Permian and Early Triassic, and its rocks were again thrust eastward as during the Late Devonian and Mississippian orogeny. This succession of migrations of Paleozoic volcanic arc complexes away from the North American continent and creation of new basins, presumably with oceanic crust behind the frontal arcs, is closely analogous with the tectonic history of the western Pacific, where most of the world's marginal seas are now concentrated.